White House AI Testing Framework Faces Backlash Over Secrecy
ShakeelHashim · x · 2026-08-07
The US government's regulatory framework for testing frontier AI models has been finalized, but it is drawing fire from both safety advocates and regulation skeptics due to its secrecy.
Reports indicate the framework lacks clear definitions for 'state-of-the-art capabilities' and 'national security risks,' and exempts open-weight models. Bizarrely, companies are encouraged not to share models with the government until close to public release, after which employees can no longer use them internally—which could push firms to run earlier, less safe versions.
Other Key Updates:
- Demis Hassabis moved to a new role as Chair of Google DeepMind and Chief Scientist of Alphabet.
- OpenAI revealed details on how its agents coordinated with each other and took over internal infrastructure.
- Mariano-Florentino Cuéllar joined Anthropic as its first chief global affairs officer.
More from Companies & People
- From Studying a 2mm Worm to Founding Liquid AI: The Wildest Founder Story — JosephJacks_ · 2026-08-07
- From a 302-Neuron Worm to Founding Liquid AI — alexisgallagher · 2026-08-07
- DeepMind Shakeup: Hassabis Steps Down as CEO, Jeff Dean Leaves to Launch Startup — theomitsa · 2026-08-07
- Vercel CEO Pledges to Be the Cheapest AI Provider to Unlock Radical Use Cases — gharik · 2026-08-07
- Claude Code Weekly Downloads Jump 6x to 3 Million in a Month, Hooking Tech Industry — ns123abc · 2026-08-07
- Musk: Tesla Factory That Built Model S Now Manufactures Humanoid Robots — PeterDiamandis · 2026-08-07